1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /* Processing of received RxRPC packets
4 * Copyright (C) 2020 Red Hat, Inc. All Rights Reserved.
5 * Written by David Howells (dhowells@redhat.com)
8 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
10 #include "ar-internal.h"
12 static void rxrpc_proto_abort(struct rxrpc_call *call, rxrpc_seq_t seq,
13 enum rxrpc_abort_reason why)
15 rxrpc_abort_call(call, seq, RX_PROTOCOL_ERROR, -EBADMSG, why);
19 * Do TCP-style congestion management [RFC 5681].
21 static void rxrpc_congestion_management(struct rxrpc_call *call,
23 struct rxrpc_ack_summary *summary,
24 rxrpc_serial_t acked_serial)
26 enum rxrpc_congest_change change = rxrpc_cong_no_change;
27 unsigned int cumulative_acks = call->cong_cumul_acks;
28 unsigned int cwnd = call->cong_cwnd;
31 summary->flight_size =
32 (call->tx_top - call->acks_hard_ack) - summary->nr_acks;
34 if (test_and_clear_bit(RXRPC_CALL_RETRANS_TIMEOUT, &call->flags)) {
35 summary->retrans_timeo = true;
36 call->cong_ssthresh = max_t(unsigned int,
37 summary->flight_size / 2, 2);
39 if (cwnd >= call->cong_ssthresh &&
40 call->cong_mode == RXRPC_CALL_SLOW_START) {
41 call->cong_mode = RXRPC_CALL_CONGEST_AVOIDANCE;
42 call->cong_tstamp = skb->tstamp;
47 cumulative_acks += summary->nr_new_acks;
48 cumulative_acks += summary->nr_rot_new_acks;
49 if (cumulative_acks > 255)
50 cumulative_acks = 255;
52 summary->mode = call->cong_mode;
53 summary->cwnd = call->cong_cwnd;
54 summary->ssthresh = call->cong_ssthresh;
55 summary->cumulative_acks = cumulative_acks;
56 summary->dup_acks = call->cong_dup_acks;
58 switch (call->cong_mode) {
59 case RXRPC_CALL_SLOW_START:
60 if (summary->saw_nacks)
61 goto packet_loss_detected;
62 if (summary->cumulative_acks > 0)
64 if (cwnd >= call->cong_ssthresh) {
65 call->cong_mode = RXRPC_CALL_CONGEST_AVOIDANCE;
66 call->cong_tstamp = skb->tstamp;
70 case RXRPC_CALL_CONGEST_AVOIDANCE:
71 if (summary->saw_nacks)
72 goto packet_loss_detected;
74 /* We analyse the number of packets that get ACK'd per RTT
75 * period and increase the window if we managed to fill it.
77 if (call->peer->rtt_count == 0)
79 if (ktime_before(skb->tstamp,
80 ktime_add_us(call->cong_tstamp,
81 call->peer->srtt_us >> 3)))
83 change = rxrpc_cong_rtt_window_end;
84 call->cong_tstamp = skb->tstamp;
85 if (cumulative_acks >= cwnd)
89 case RXRPC_CALL_PACKET_LOSS:
90 if (!summary->saw_nacks)
91 goto resume_normality;
93 if (summary->new_low_nack) {
94 change = rxrpc_cong_new_low_nack;
95 call->cong_dup_acks = 1;
96 if (call->cong_extra > 1)
101 call->cong_dup_acks++;
102 if (call->cong_dup_acks < 3)
103 goto send_extra_data;
105 change = rxrpc_cong_begin_retransmission;
106 call->cong_mode = RXRPC_CALL_FAST_RETRANSMIT;
107 call->cong_ssthresh = max_t(unsigned int,
108 summary->flight_size / 2, 2);
109 cwnd = call->cong_ssthresh + 3;
110 call->cong_extra = 0;
111 call->cong_dup_acks = 0;
115 case RXRPC_CALL_FAST_RETRANSMIT:
116 if (!summary->new_low_nack) {
117 if (summary->nr_new_acks == 0)
119 call->cong_dup_acks++;
120 if (call->cong_dup_acks == 2) {
121 change = rxrpc_cong_retransmit_again;
122 call->cong_dup_acks = 0;
126 change = rxrpc_cong_progress;
127 cwnd = call->cong_ssthresh;
128 if (!summary->saw_nacks)
129 goto resume_normality;
139 change = rxrpc_cong_cleared_nacks;
140 call->cong_dup_acks = 0;
141 call->cong_extra = 0;
142 call->cong_tstamp = skb->tstamp;
143 if (cwnd < call->cong_ssthresh)
144 call->cong_mode = RXRPC_CALL_SLOW_START;
146 call->cong_mode = RXRPC_CALL_CONGEST_AVOIDANCE;
150 if (cwnd >= RXRPC_TX_MAX_WINDOW)
151 cwnd = RXRPC_TX_MAX_WINDOW;
152 call->cong_cwnd = cwnd;
153 call->cong_cumul_acks = cumulative_acks;
154 trace_rxrpc_congest(call, summary, acked_serial, change);
156 rxrpc_resend(call, skb);
159 packet_loss_detected:
160 change = rxrpc_cong_saw_nack;
161 call->cong_mode = RXRPC_CALL_PACKET_LOSS;
162 call->cong_dup_acks = 0;
163 goto send_extra_data;
166 /* Send some previously unsent DATA if we have some to advance the ACK
169 if (test_bit(RXRPC_CALL_TX_LAST, &call->flags) ||
170 summary->nr_acks != call->tx_top - call->acks_hard_ack) {
172 wake_up(&call->waitq);
174 goto out_no_clear_ca;
178 * Degrade the congestion window if we haven't transmitted a packet for >1RTT.
180 void rxrpc_congestion_degrade(struct rxrpc_call *call)
184 if (call->cong_mode != RXRPC_CALL_SLOW_START &&
185 call->cong_mode != RXRPC_CALL_CONGEST_AVOIDANCE)
187 if (__rxrpc_call_state(call) == RXRPC_CALL_CLIENT_AWAIT_REPLY)
190 rtt = ns_to_ktime(call->peer->srtt_us * (1000 / 8));
191 now = ktime_get_real();
192 if (!ktime_before(ktime_add(call->tx_last_sent, rtt), now))
195 trace_rxrpc_reset_cwnd(call, now);
196 rxrpc_inc_stat(call->rxnet, stat_tx_data_cwnd_reset);
197 call->tx_last_sent = now;
198 call->cong_mode = RXRPC_CALL_SLOW_START;
199 call->cong_ssthresh = max_t(unsigned int, call->cong_ssthresh,
200 call->cong_cwnd * 3 / 4);
201 call->cong_cwnd = max_t(unsigned int, call->cong_cwnd / 2, RXRPC_MIN_CWND);
205 * Apply a hard ACK by advancing the Tx window.
207 static bool rxrpc_rotate_tx_window(struct rxrpc_call *call, rxrpc_seq_t to,
208 struct rxrpc_ack_summary *summary)
210 struct rxrpc_txbuf *txb;
211 bool rot_last = false;
213 list_for_each_entry_rcu(txb, &call->tx_buffer, call_link, false) {
214 if (before_eq(txb->seq, call->acks_hard_ack))
216 summary->nr_rot_new_acks++;
217 if (test_bit(RXRPC_TXBUF_LAST, &txb->flags)) {
218 set_bit(RXRPC_CALL_TX_LAST, &call->flags);
226 set_bit(RXRPC_CALL_TX_ALL_ACKED, &call->flags);
228 _enter("%x,%x,%x,%d", to, call->acks_hard_ack, call->tx_top, rot_last);
230 if (call->acks_lowest_nak == call->acks_hard_ack) {
231 call->acks_lowest_nak = to;
232 } else if (after(to, call->acks_lowest_nak)) {
233 summary->new_low_nack = true;
234 call->acks_lowest_nak = to;
237 smp_store_release(&call->acks_hard_ack, to);
239 trace_rxrpc_txqueue(call, (rot_last ?
240 rxrpc_txqueue_rotate_last :
241 rxrpc_txqueue_rotate));
242 wake_up(&call->waitq);
247 * End the transmission phase of a call.
249 * This occurs when we get an ACKALL packet, the first DATA packet of a reply,
250 * or a final ACK packet.
252 static void rxrpc_end_tx_phase(struct rxrpc_call *call, bool reply_begun,
253 enum rxrpc_abort_reason abort_why)
255 ASSERT(test_bit(RXRPC_CALL_TX_LAST, &call->flags));
257 switch (__rxrpc_call_state(call)) {
258 case RXRPC_CALL_CLIENT_SEND_REQUEST:
259 case RXRPC_CALL_CLIENT_AWAIT_REPLY:
261 rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_RECV_REPLY);
262 trace_rxrpc_txqueue(call, rxrpc_txqueue_end);
266 rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_AWAIT_REPLY);
267 trace_rxrpc_txqueue(call, rxrpc_txqueue_await_reply);
270 case RXRPC_CALL_SERVER_AWAIT_ACK:
271 rxrpc_call_completed(call);
272 trace_rxrpc_txqueue(call, rxrpc_txqueue_end);
276 kdebug("end_tx %s", rxrpc_call_states[__rxrpc_call_state(call)]);
277 rxrpc_proto_abort(call, call->tx_top, abort_why);
283 * Begin the reply reception phase of a call.
285 static bool rxrpc_receiving_reply(struct rxrpc_call *call)
287 struct rxrpc_ack_summary summary = { 0 };
288 unsigned long now, timo;
289 rxrpc_seq_t top = READ_ONCE(call->tx_top);
291 if (call->ackr_reason) {
293 timo = now + MAX_JIFFY_OFFSET;
295 WRITE_ONCE(call->delay_ack_at, timo);
296 trace_rxrpc_timer(call, rxrpc_timer_init_for_reply, now);
299 if (!test_bit(RXRPC_CALL_TX_LAST, &call->flags)) {
300 if (!rxrpc_rotate_tx_window(call, top, &summary)) {
301 rxrpc_proto_abort(call, top, rxrpc_eproto_early_reply);
306 rxrpc_end_tx_phase(call, true, rxrpc_eproto_unexpected_reply);
311 * End the packet reception phase.
313 static void rxrpc_end_rx_phase(struct rxrpc_call *call, rxrpc_serial_t serial)
315 rxrpc_seq_t whigh = READ_ONCE(call->rx_highest_seq);
317 _enter("%d,%s", call->debug_id, rxrpc_call_states[__rxrpc_call_state(call)]);
319 trace_rxrpc_receive(call, rxrpc_receive_end, 0, whigh);
321 switch (__rxrpc_call_state(call)) {
322 case RXRPC_CALL_CLIENT_RECV_REPLY:
323 rxrpc_propose_delay_ACK(call, serial, rxrpc_propose_ack_terminal_ack);
324 rxrpc_call_completed(call);
327 case RXRPC_CALL_SERVER_RECV_REQUEST:
328 rxrpc_set_call_state(call, RXRPC_CALL_SERVER_ACK_REQUEST);
329 call->expect_req_by = jiffies + MAX_JIFFY_OFFSET;
330 rxrpc_propose_delay_ACK(call, serial, rxrpc_propose_ack_processing_op);
338 static void rxrpc_input_update_ack_window(struct rxrpc_call *call,
339 rxrpc_seq_t window, rxrpc_seq_t wtop)
341 call->ackr_window = window;
342 call->ackr_wtop = wtop;
346 * Push a DATA packet onto the Rx queue.
348 static void rxrpc_input_queue_data(struct rxrpc_call *call, struct sk_buff *skb,
349 rxrpc_seq_t window, rxrpc_seq_t wtop,
350 enum rxrpc_receive_trace why)
352 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
353 bool last = sp->hdr.flags & RXRPC_LAST_PACKET;
355 __skb_queue_tail(&call->recvmsg_queue, skb);
356 rxrpc_input_update_ack_window(call, window, wtop);
357 trace_rxrpc_receive(call, last ? why + 1 : why, sp->hdr.serial, sp->hdr.seq);
359 rxrpc_end_rx_phase(call, sp->hdr.serial);
363 * Process a DATA packet.
365 static void rxrpc_input_data_one(struct rxrpc_call *call, struct sk_buff *skb,
368 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
370 rxrpc_serial_t serial = sp->hdr.serial;
371 unsigned int sack = call->ackr_sack_base;
372 rxrpc_seq_t window = call->ackr_window;
373 rxrpc_seq_t wtop = call->ackr_wtop;
374 rxrpc_seq_t wlimit = window + call->rx_winsize - 1;
375 rxrpc_seq_t seq = sp->hdr.seq;
376 bool last = sp->hdr.flags & RXRPC_LAST_PACKET;
379 rxrpc_inc_stat(call->rxnet, stat_rx_data);
380 if (sp->hdr.flags & RXRPC_REQUEST_ACK)
381 rxrpc_inc_stat(call->rxnet, stat_rx_data_reqack);
382 if (sp->hdr.flags & RXRPC_JUMBO_PACKET)
383 rxrpc_inc_stat(call->rxnet, stat_rx_data_jumbo);
386 if (test_and_set_bit(RXRPC_CALL_RX_LAST, &call->flags) &&
388 return rxrpc_proto_abort(call, seq, rxrpc_eproto_different_last);
390 if (test_bit(RXRPC_CALL_RX_LAST, &call->flags) &&
391 after_eq(seq, wtop)) {
392 pr_warn("Packet beyond last: c=%x q=%x window=%x-%x wlimit=%x\n",
393 call->debug_id, seq, window, wtop, wlimit);
394 return rxrpc_proto_abort(call, seq, rxrpc_eproto_data_after_last);
398 if (after(seq, call->rx_highest_seq))
399 call->rx_highest_seq = seq;
401 trace_rxrpc_rx_data(call->debug_id, seq, serial, sp->hdr.flags);
403 if (before(seq, window)) {
404 ack_reason = RXRPC_ACK_DUPLICATE;
407 if (after(seq, wlimit)) {
408 ack_reason = RXRPC_ACK_EXCEEDS_WINDOW;
412 /* Queue the packet. */
414 if (sp->hdr.flags & RXRPC_REQUEST_ACK)
415 ack_reason = RXRPC_ACK_REQUESTED;
416 /* Send an immediate ACK if we fill in a hole */
417 else if (!skb_queue_empty(&call->rx_oos_queue))
418 ack_reason = RXRPC_ACK_DELAY;
420 call->ackr_nr_unacked++;
423 if (after(window, wtop)) {
424 trace_rxrpc_sack(call, seq, sack, rxrpc_sack_none);
427 trace_rxrpc_sack(call, seq, sack, rxrpc_sack_advance);
428 sack = (sack + 1) % RXRPC_SACK_SIZE;
432 rxrpc_get_skb(skb, rxrpc_skb_get_to_recvmsg);
434 spin_lock(&call->recvmsg_queue.lock);
435 rxrpc_input_queue_data(call, skb, window, wtop, rxrpc_receive_queue);
438 while ((oos = skb_peek(&call->rx_oos_queue))) {
439 struct rxrpc_skb_priv *osp = rxrpc_skb(oos);
441 if (after(osp->hdr.seq, window))
444 __skb_unlink(oos, &call->rx_oos_queue);
445 last = osp->hdr.flags & RXRPC_LAST_PACKET;
447 call->ackr_sack_table[sack] = 0;
448 trace_rxrpc_sack(call, seq, sack, rxrpc_sack_fill);
449 sack = (sack + 1) % RXRPC_SACK_SIZE;
452 rxrpc_input_queue_data(call, oos, window, wtop,
453 rxrpc_receive_queue_oos);
456 spin_unlock(&call->recvmsg_queue.lock);
458 call->ackr_sack_base = sack;
462 ack_reason = RXRPC_ACK_OUT_OF_SEQUENCE;
465 sack = (sack + slot) % RXRPC_SACK_SIZE;
467 if (call->ackr_sack_table[sack % RXRPC_SACK_SIZE]) {
468 ack_reason = RXRPC_ACK_DUPLICATE;
472 call->ackr_sack_table[sack % RXRPC_SACK_SIZE] |= 1;
473 trace_rxrpc_sack(call, seq, sack, rxrpc_sack_oos);
475 if (after(seq + 1, wtop)) {
477 rxrpc_input_update_ack_window(call, window, wtop);
480 skb_queue_walk(&call->rx_oos_queue, oos) {
481 struct rxrpc_skb_priv *osp = rxrpc_skb(oos);
483 if (after(osp->hdr.seq, seq)) {
484 rxrpc_get_skb(skb, rxrpc_skb_get_to_recvmsg_oos);
485 __skb_queue_before(&call->rx_oos_queue, oos, skb);
490 rxrpc_get_skb(skb, rxrpc_skb_get_to_recvmsg_oos);
491 __skb_queue_tail(&call->rx_oos_queue, skb);
493 trace_rxrpc_receive(call, last ? rxrpc_receive_oos_last : rxrpc_receive_oos,
494 sp->hdr.serial, sp->hdr.seq);
499 rxrpc_send_ACK(call, ack_reason, serial,
500 rxrpc_propose_ack_input_data);
502 rxrpc_propose_delay_ACK(call, serial,
503 rxrpc_propose_ack_input_data);
507 * Split a jumbo packet and file the bits separately.
509 static bool rxrpc_input_split_jumbo(struct rxrpc_call *call, struct sk_buff *skb)
511 struct rxrpc_jumbo_header jhdr;
512 struct rxrpc_skb_priv *sp = rxrpc_skb(skb), *jsp;
513 struct sk_buff *jskb;
514 unsigned int offset = sizeof(struct rxrpc_wire_header);
515 unsigned int len = skb->len - offset;
518 while (sp->hdr.flags & RXRPC_JUMBO_PACKET) {
519 if (len < RXRPC_JUMBO_SUBPKTLEN)
521 if (sp->hdr.flags & RXRPC_LAST_PACKET)
523 if (skb_copy_bits(skb, offset + RXRPC_JUMBO_DATALEN,
524 &jhdr, sizeof(jhdr)) < 0)
527 jskb = skb_clone(skb, GFP_NOFS);
529 kdebug("couldn't clone");
532 rxrpc_new_skb(jskb, rxrpc_skb_new_jumbo_subpacket);
533 jsp = rxrpc_skb(jskb);
534 jsp->offset = offset;
535 jsp->len = RXRPC_JUMBO_DATALEN;
536 rxrpc_input_data_one(call, jskb, ¬ify);
537 rxrpc_free_skb(jskb, rxrpc_skb_put_jumbo_subpacket);
539 sp->hdr.flags = jhdr.flags;
540 sp->hdr._rsvd = ntohs(jhdr._rsvd);
543 offset += RXRPC_JUMBO_SUBPKTLEN;
544 len -= RXRPC_JUMBO_SUBPKTLEN;
549 rxrpc_input_data_one(call, skb, ¬ify);
551 trace_rxrpc_notify_socket(call->debug_id, sp->hdr.serial);
552 rxrpc_notify_socket(call);
561 * Process a DATA packet, adding the packet to the Rx ring. The caller's
562 * packet ref must be passed on or discarded.
564 static void rxrpc_input_data(struct rxrpc_call *call, struct sk_buff *skb)
566 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
567 rxrpc_serial_t serial = sp->hdr.serial;
568 rxrpc_seq_t seq0 = sp->hdr.seq;
570 _enter("{%x,%x,%x},{%u,%x}",
571 call->ackr_window, call->ackr_wtop, call->rx_highest_seq,
574 if (__rxrpc_call_is_complete(call))
577 switch (__rxrpc_call_state(call)) {
578 case RXRPC_CALL_CLIENT_SEND_REQUEST:
579 case RXRPC_CALL_CLIENT_AWAIT_REPLY:
580 /* Received data implicitly ACKs all of the request
581 * packets we sent when we're acting as a client.
583 if (!rxrpc_receiving_reply(call))
587 case RXRPC_CALL_SERVER_RECV_REQUEST: {
588 unsigned long timo = READ_ONCE(call->next_req_timo);
589 unsigned long now, expect_req_by;
593 expect_req_by = now + timo;
594 WRITE_ONCE(call->expect_req_by, expect_req_by);
595 rxrpc_reduce_call_timer(call, expect_req_by, now,
596 rxrpc_timer_set_for_idle);
605 if (!rxrpc_input_split_jumbo(call, skb)) {
606 rxrpc_proto_abort(call, sp->hdr.seq, rxrpc_badmsg_bad_jumbo);
612 trace_rxrpc_notify_socket(call->debug_id, serial);
613 rxrpc_notify_socket(call);
618 * See if there's a cached RTT probe to complete.
620 static void rxrpc_complete_rtt_probe(struct rxrpc_call *call,
622 rxrpc_serial_t acked_serial,
623 rxrpc_serial_t ack_serial,
624 enum rxrpc_rtt_rx_trace type)
626 rxrpc_serial_t orig_serial;
629 bool matched = false;
632 avail = READ_ONCE(call->rtt_avail);
633 smp_rmb(); /* Read avail bits before accessing data. */
635 for (i = 0; i < ARRAY_SIZE(call->rtt_serial); i++) {
636 if (!test_bit(i + RXRPC_CALL_RTT_PEND_SHIFT, &avail))
639 sent_at = call->rtt_sent_at[i];
640 orig_serial = call->rtt_serial[i];
642 if (orig_serial == acked_serial) {
643 clear_bit(i + RXRPC_CALL_RTT_PEND_SHIFT, &call->rtt_avail);
644 smp_mb(); /* Read data before setting avail bit */
645 set_bit(i, &call->rtt_avail);
646 if (type != rxrpc_rtt_rx_cancel)
647 rxrpc_peer_add_rtt(call, type, i, acked_serial, ack_serial,
650 trace_rxrpc_rtt_rx(call, rxrpc_rtt_rx_cancel, i,
651 orig_serial, acked_serial, 0, 0);
655 /* If a later serial is being acked, then mark this slot as
658 if (after(acked_serial, orig_serial)) {
659 trace_rxrpc_rtt_rx(call, rxrpc_rtt_rx_obsolete, i,
660 orig_serial, acked_serial, 0, 0);
661 clear_bit(i + RXRPC_CALL_RTT_PEND_SHIFT, &call->rtt_avail);
663 set_bit(i, &call->rtt_avail);
668 trace_rxrpc_rtt_rx(call, rxrpc_rtt_rx_lost, 9, 0, acked_serial, 0, 0);
672 * Process the extra information that may be appended to an ACK packet
674 static void rxrpc_input_ackinfo(struct rxrpc_call *call, struct sk_buff *skb,
675 struct rxrpc_ackinfo *ackinfo)
677 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
678 struct rxrpc_peer *peer;
681 u32 rwind = ntohl(ackinfo->rwind);
683 if (rwind > RXRPC_TX_MAX_WINDOW)
684 rwind = RXRPC_TX_MAX_WINDOW;
685 if (call->tx_winsize != rwind) {
686 if (rwind > call->tx_winsize)
688 trace_rxrpc_rx_rwind_change(call, sp->hdr.serial, rwind, wake);
689 call->tx_winsize = rwind;
692 if (call->cong_ssthresh > rwind)
693 call->cong_ssthresh = rwind;
695 mtu = min(ntohl(ackinfo->rxMTU), ntohl(ackinfo->maxMTU));
698 if (mtu < peer->maxdata) {
699 spin_lock(&peer->lock);
701 peer->mtu = mtu + peer->hdrsize;
702 spin_unlock(&peer->lock);
706 wake_up(&call->waitq);
710 * Process individual soft ACKs.
712 * Each ACK in the array corresponds to one packet and can be either an ACK or
713 * a NAK. If we get find an explicitly NAK'd packet we resend immediately;
714 * packets that lie beyond the end of the ACK list are scheduled for resend by
715 * the timer on the basis that the peer might just not have processed them at
716 * the time the ACK was sent.
718 static void rxrpc_input_soft_acks(struct rxrpc_call *call, u8 *acks,
719 rxrpc_seq_t seq, int nr_acks,
720 struct rxrpc_ack_summary *summary)
724 for (i = 0; i < nr_acks; i++) {
725 if (acks[i] == RXRPC_ACK_TYPE_ACK) {
727 summary->nr_new_acks++;
729 if (!summary->saw_nacks &&
730 call->acks_lowest_nak != seq + i) {
731 call->acks_lowest_nak = seq + i;
732 summary->new_low_nack = true;
734 summary->saw_nacks = true;
740 * Return true if the ACK is valid - ie. it doesn't appear to have regressed
741 * with respect to the ack state conveyed by preceding ACKs.
743 static bool rxrpc_is_ack_valid(struct rxrpc_call *call,
744 rxrpc_seq_t first_pkt, rxrpc_seq_t prev_pkt)
746 rxrpc_seq_t base = READ_ONCE(call->acks_first_seq);
748 if (after(first_pkt, base))
749 return true; /* The window advanced */
751 if (before(first_pkt, base))
752 return false; /* firstPacket regressed */
754 if (after_eq(prev_pkt, call->acks_prev_seq))
755 return true; /* previousPacket hasn't regressed. */
757 /* Some rx implementations put a serial number in previousPacket. */
758 if (after_eq(prev_pkt, base + call->tx_winsize))
764 * Process an ACK packet.
766 * ack.firstPacket is the sequence number of the first soft-ACK'd/NAK'd packet
767 * in the ACK array. Anything before that is hard-ACK'd and may be discarded.
769 * A hard-ACK means that a packet has been processed and may be discarded; a
770 * soft-ACK means that the packet may be discarded and retransmission
771 * requested. A phase is complete when all packets are hard-ACK'd.
773 static void rxrpc_input_ack(struct rxrpc_call *call, struct sk_buff *skb)
775 struct rxrpc_ack_summary summary = { 0 };
776 struct rxrpc_ackpacket ack;
777 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
778 struct rxrpc_ackinfo info;
779 rxrpc_serial_t ack_serial, acked_serial;
780 rxrpc_seq_t first_soft_ack, hard_ack, prev_pkt;
781 int nr_acks, offset, ioffset;
785 offset = sizeof(struct rxrpc_wire_header);
786 if (skb_copy_bits(skb, offset, &ack, sizeof(ack)) < 0)
787 return rxrpc_proto_abort(call, 0, rxrpc_badmsg_short_ack);
788 offset += sizeof(ack);
790 ack_serial = sp->hdr.serial;
791 acked_serial = ntohl(ack.serial);
792 first_soft_ack = ntohl(ack.firstPacket);
793 prev_pkt = ntohl(ack.previousPacket);
794 hard_ack = first_soft_ack - 1;
796 summary.ack_reason = (ack.reason < RXRPC_ACK__INVALID ?
797 ack.reason : RXRPC_ACK__INVALID);
799 trace_rxrpc_rx_ack(call, ack_serial, acked_serial,
800 first_soft_ack, prev_pkt,
801 summary.ack_reason, nr_acks);
802 rxrpc_inc_stat(call->rxnet, stat_rx_acks[ack.reason]);
804 switch (ack.reason) {
805 case RXRPC_ACK_PING_RESPONSE:
806 rxrpc_complete_rtt_probe(call, skb->tstamp, acked_serial, ack_serial,
807 rxrpc_rtt_rx_ping_response);
809 case RXRPC_ACK_REQUESTED:
810 rxrpc_complete_rtt_probe(call, skb->tstamp, acked_serial, ack_serial,
811 rxrpc_rtt_rx_requested_ack);
814 if (acked_serial != 0)
815 rxrpc_complete_rtt_probe(call, skb->tstamp, acked_serial, ack_serial,
816 rxrpc_rtt_rx_cancel);
820 if (ack.reason == RXRPC_ACK_PING) {
821 rxrpc_send_ACK(call, RXRPC_ACK_PING_RESPONSE, ack_serial,
822 rxrpc_propose_ack_respond_to_ping);
823 } else if (sp->hdr.flags & RXRPC_REQUEST_ACK) {
824 rxrpc_send_ACK(call, RXRPC_ACK_REQUESTED, ack_serial,
825 rxrpc_propose_ack_respond_to_ack);
828 /* If we get an EXCEEDS_WINDOW ACK from the server, it probably
829 * indicates that the client address changed due to NAT. The server
830 * lost the call because it switched to a different peer.
832 if (unlikely(ack.reason == RXRPC_ACK_EXCEEDS_WINDOW) &&
833 first_soft_ack == 1 &&
835 rxrpc_is_client_call(call)) {
836 rxrpc_set_call_completion(call, RXRPC_CALL_REMOTELY_ABORTED,
841 /* If we get an OUT_OF_SEQUENCE ACK from the server, that can also
842 * indicate a change of address. However, we can retransmit the call
843 * if we still have it buffered to the beginning.
845 if (unlikely(ack.reason == RXRPC_ACK_OUT_OF_SEQUENCE) &&
846 first_soft_ack == 1 &&
848 call->acks_hard_ack == 0 &&
849 rxrpc_is_client_call(call)) {
850 rxrpc_set_call_completion(call, RXRPC_CALL_REMOTELY_ABORTED,
855 /* Discard any out-of-order or duplicate ACKs (outside lock). */
856 if (!rxrpc_is_ack_valid(call, first_soft_ack, prev_pkt)) {
857 trace_rxrpc_rx_discard_ack(call->debug_id, ack_serial,
858 first_soft_ack, call->acks_first_seq,
859 prev_pkt, call->acks_prev_seq);
864 ioffset = offset + nr_acks + 3;
865 if (skb->len >= ioffset + sizeof(info) &&
866 skb_copy_bits(skb, ioffset, &info, sizeof(info)) < 0)
867 return rxrpc_proto_abort(call, 0, rxrpc_badmsg_short_ack_info);
872 call->acks_latest_ts = skb->tstamp;
873 call->acks_first_seq = first_soft_ack;
874 call->acks_prev_seq = prev_pkt;
876 switch (ack.reason) {
880 if (after(acked_serial, call->acks_highest_serial))
881 call->acks_highest_serial = acked_serial;
885 /* Parse rwind and mtu sizes if provided. */
887 rxrpc_input_ackinfo(call, skb, &info);
889 if (first_soft_ack == 0)
890 return rxrpc_proto_abort(call, 0, rxrpc_eproto_ackr_zero);
892 /* Ignore ACKs unless we are or have just been transmitting. */
893 switch (__rxrpc_call_state(call)) {
894 case RXRPC_CALL_CLIENT_SEND_REQUEST:
895 case RXRPC_CALL_CLIENT_AWAIT_REPLY:
896 case RXRPC_CALL_SERVER_SEND_REPLY:
897 case RXRPC_CALL_SERVER_AWAIT_ACK:
903 if (before(hard_ack, call->acks_hard_ack) ||
904 after(hard_ack, call->tx_top))
905 return rxrpc_proto_abort(call, 0, rxrpc_eproto_ackr_outside_window);
906 if (nr_acks > call->tx_top - hard_ack)
907 return rxrpc_proto_abort(call, 0, rxrpc_eproto_ackr_sack_overflow);
909 if (after(hard_ack, call->acks_hard_ack)) {
910 if (rxrpc_rotate_tx_window(call, hard_ack, &summary)) {
911 rxrpc_end_tx_phase(call, false, rxrpc_eproto_unexpected_ack);
917 if (offset > (int)skb->len - nr_acks)
918 return rxrpc_proto_abort(call, 0, rxrpc_eproto_ackr_short_sack);
919 rxrpc_input_soft_acks(call, skb->data + offset, first_soft_ack,
923 if (test_bit(RXRPC_CALL_TX_LAST, &call->flags) &&
924 summary.nr_acks == call->tx_top - hard_ack &&
925 rxrpc_is_client_call(call))
926 rxrpc_propose_ping(call, ack_serial,
927 rxrpc_propose_ack_ping_for_lost_reply);
929 rxrpc_congestion_management(call, skb, &summary, acked_serial);
933 * Process an ACKALL packet.
935 static void rxrpc_input_ackall(struct rxrpc_call *call, struct sk_buff *skb)
937 struct rxrpc_ack_summary summary = { 0 };
939 if (rxrpc_rotate_tx_window(call, call->tx_top, &summary))
940 rxrpc_end_tx_phase(call, false, rxrpc_eproto_unexpected_ackall);
944 * Process an ABORT packet directed at a call.
946 static void rxrpc_input_abort(struct rxrpc_call *call, struct sk_buff *skb)
948 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
950 trace_rxrpc_rx_abort(call, sp->hdr.serial, skb->priority);
952 rxrpc_set_call_completion(call, RXRPC_CALL_REMOTELY_ABORTED,
953 skb->priority, -ECONNABORTED);
957 * Process an incoming call packet.
959 void rxrpc_input_call_packet(struct rxrpc_call *call, struct sk_buff *skb)
961 struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
964 _enter("%p,%p", call, skb);
966 if (sp->hdr.serviceId != call->dest_srx.srx_service)
967 call->dest_srx.srx_service = sp->hdr.serviceId;
968 if ((int)sp->hdr.serial - (int)call->rx_serial > 0)
969 call->rx_serial = sp->hdr.serial;
970 if (!test_bit(RXRPC_CALL_RX_HEARD, &call->flags))
971 set_bit(RXRPC_CALL_RX_HEARD, &call->flags);
973 timo = READ_ONCE(call->next_rx_timo);
975 unsigned long now = jiffies, expect_rx_by;
977 expect_rx_by = now + timo;
978 WRITE_ONCE(call->expect_rx_by, expect_rx_by);
979 rxrpc_reduce_call_timer(call, expect_rx_by, now,
980 rxrpc_timer_set_for_normal);
983 switch (sp->hdr.type) {
984 case RXRPC_PACKET_TYPE_DATA:
985 return rxrpc_input_data(call, skb);
987 case RXRPC_PACKET_TYPE_ACK:
988 return rxrpc_input_ack(call, skb);
990 case RXRPC_PACKET_TYPE_BUSY:
991 /* Just ignore BUSY packets from the server; the retry and
992 * lifespan timers will take care of business. BUSY packets
993 * from the client don't make sense.
997 case RXRPC_PACKET_TYPE_ABORT:
998 return rxrpc_input_abort(call, skb);
1000 case RXRPC_PACKET_TYPE_ACKALL:
1001 return rxrpc_input_ackall(call, skb);
1009 * Handle a new service call on a channel implicitly completing the preceding
1010 * call on that channel. This does not apply to client conns.
1012 * TODO: If callNumber > call_id + 1, renegotiate security.
1014 void rxrpc_implicit_end_call(struct rxrpc_call *call, struct sk_buff *skb)
1016 switch (__rxrpc_call_state(call)) {
1017 case RXRPC_CALL_SERVER_AWAIT_ACK:
1018 rxrpc_call_completed(call);
1020 case RXRPC_CALL_COMPLETE:
1023 rxrpc_abort_call(call, 0, RX_CALL_DEAD, -ESHUTDOWN,
1024 rxrpc_eproto_improper_term);
1025 trace_rxrpc_improper_term(call);
1029 rxrpc_input_call_event(call, skb);